output gap stationarity

[Viktoria]

Is it right to take the 1st difference of the "output gap" data? It is already a difference (difference between the actual output of an economy and its potential output) but the data series is not stationary. I have calculated the potential output by using HP filter. But I did not correct for stationarity in my initial output dara series. Could you please advise me how to correct for this?

[trubador]

Nonstationarity is not a problem by itself. It depends on the context or the method you use. And it may arise due to a number of reasons. Since you have already detrended the series, existence of outlier(s) or structural break(s) are the usual suspects here. Nevertheless, HP filter may have done a poor job and you may need to employ a different filter or simply change the value of penalty parameter.

[Viktoria]

Thank you very much Trubador for the reply! But I still can not understand is it wrong to take the 1st difference of output gap (that is d(output_gap)) ?

[trubador]

It depends on the context or the method you use.